Detection of process-related changes in plant patterns at extended spatial scales during early dryland desertification

Abstract

Arid and semiarid shrublands occupy extensive land areas over the world, are susceptible to desertification by anthropic use and can contribute to regional climate change. These prompt the interest to monitor and evaluate these lands adequately in order to detect early stages of degradation. Evaluation topics must refer to biologyrelevant characteristics of these systems, while simultaneously satisfying sampling consistency over extended landscape areas. We present an analysis of process-relevant parameters related to changes in the spatial arrangement of the plant canopy of shrublands inferred from high-resolution panchromatic aerial photos and Interferometric Synthetic Aperture Radar imagery. We obtained low-altitude images systematically located along several gradients of land-use intensity in a Patagonian Monte shrubland in Argentina. Images were digitized to spatial resolutions ranging from 0.09 to 0.72 m (pixel size) and the average values and an-isotropic characteristics of the plant canopy patterns were quantified by means of a Fourier metric. We used radar-derived imagery to overlay the panchromatic images on a digital elevation model in order to study the correspondence of potential runoff patterns and the spatial arrangement of plants. We related an-isotropic features of the plant canopy images to the prevailing wind regime. Observed trends were further interpreted on the basis of a spatial-explicit simulation model describing the dynamics of the main functional groups in the plant community. We conclude that early stages of anthropic-driven dryland degradation in the Patagonian Monte can be characterized by the incipient un-coupling of spatial vegetation patterns from those of runoff at a landscape scale, and a progressive coupling to the spatial pattern of the wind regime. The method and metrics we present can be used to quantify early desertification changes in other similar drylands at extended spatial scales.